Application Research Of Track Detection System Based On Cell Phone Signaling

With the development of mobile communication technology and the popularization of mobile terminals such as smart phones,the communication data shows exponential growth.Mobile phone users have frequent services,including telephone services,SMS services,Internet services,location update services,etc.A large number of users and high-frequency communication behaviors have generated massive amounts of mobile phone data.This massive amount of information cannot be analyzed manually.However,with the advent of big data technology,the value of these data has become increasingly prominent.In this context,this thesis proposes a method of target trajectory analysis and extraction based on mobile phone signaling big data.At present,GPS positioning data is mainly used for trajectory analysis research,which has high positioning accuracy but is greatly affected by weather and other factors.Although the accuracy of mobile phone signaling data is not as good as GPS,it has the advantages of large data volume,wide coverage,and low acquisition cost,it is an ideal and high-quality data source.This thesis designs and implements a big data platform system for the analysis and visualization of target trajectories Based on the mobile phone signaling information data.First,I conducted indepth study and research on big data technology and theory,and elaborated the process of business analysis and system architecture design,including the design and implementation of various functional modules,mainly including data collection,cleaning,storage,analysis and visualization,etc.Map Reduce is a distributed computing framework provided by Hadoop,but the computing speed is slow due to its own limitations,and spark uses a memory computing model,which makes the processing speed much faster than Map Reduce.Therefore,this thesis uses Spark as the analysis engine of the big data platform.Aiming at the problem of data skew in Spark big data analysis,this thesis proposes an optimization scheme that adds random prefixes to make the data of each node more uniform,thereby shortening the execution time of the task.At the same time,a method based on cost model optimization is proposed in the SQL associated query.Finally,relevant experiments are designed to verify the effectiveness of the scheme proposed in this thesis,and good results have been achieved.